On Fri, Sep 06, 2002 at 12:39:02AM +0200, Ingo Molnar wrote:
>
> On Thu, 5 Sep 2002, Daniel Jacobowitz wrote:
>
> > If we want to do this then we'd need to fix up every ptrace
> > implementation in every architecture to call the appropriate function;
> > it's a separate problem.
>
> which code relies on having debugged children only in the ->children list
> and not in the ->ptrace_children list?
Every implementation of PTRACE_TRACEME leaves them in the ->children
list. They are never added to ptrace_children. Whether this is
_right_ is another question.
> > > i'm not sure about this either. What happens if an (untraced) parent has
> > > traced and untraced children, and does a wait4. Would it confuse the
> > > debugger if the parent could get one of the traced tasks as a result in
> > > wait4? And how does the debugger solve this problem?
> >
> > Well, it seems to me that when a traced task has an event, it should be
> > reported first to the debugger - for signals this happens in do_signal -
> > and then possibly to the normal parent. But I'm not sure if this
> > actually happens right now or not. Worth investigating some more.
>
> it just cannot happen. There are only two kinds of events passed via
> wait4: tracing related and exit related. An exiting task is not traced
> anymore. And two tasks cannot trace the same task - so it can never happen
> that wait4 wants to look at ->ptrace_children for events.
Oh. You are, of course, right.
-- Daniel Jacobowitz MontaVista Software Debian GNU/Linux Developer - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Sep 07 2002 - 22:00:27 EST